Rockmass reaction to blasting in mines studied by local induced seismicity

Elaborate manual processing of mining-induced seismic events following production blasting in two of the oldest underground Swedish mines (Zinkgruvan and Garpenberg) was carried out in order to obtain the most accurate possible solutions for kinematic and dynamic parameters and to obtain information about the reaction of the rock mass to blasting. The main problems for each mine were not the same so slightly different procedures were followed. Data used in the thesis was recorded by small scale temporary local seismic systems, consisting of 16 to 18 sensors, installed in both mines in 2015 to study the local site effects and ground motion from possible damaging seismic events. Additional data from the permanent mine-wide seismic system in Garpenberg was also used. A rockburst event occurred on July 2, 2015 (Mw 1.2) at 15:01 during a blast at Zinkgruvan Mine that caused damage. The source parameters of the seismic events following the blast on July 2 were studied and compared with the parameters of the events following another blast on July 1 (in total 4517 waveforms from 271 events were analyzed). The moment magnitude of all events ranged between -3.3 and 1.2. The events on July 1 and July 2 formed two separate large clusters (with length up to 60 to 80 m) with orientations of long axes perpendicular to each other. The largest moment magnitude events in each cluster were located at the edges of the corresponding clusters: Mw 0.6 event on July 1 was on the top of the first cluster, however, the Mw 1.2 event on July 2 was on the bottom of the other cluster. Dynamic source parameters of seismic events on July 1 and 2 showed consistently different values and relationships. In general the events on July 1 had smaller magnitudes, seismic moment, seismic energy, apparent stress and stress drop compared with the events on July 2. The source parameters relationships with each other were also examined and found different. The second case study was related to seismicity caused by production blasting in Garpenberg Mine. The parameters of the events within the period October 8 to 13, 2015 were studied. Three different datasets, for the same 30 events recorded by the permanent seismic system, temporary local seismic system, and merged data from both systems, were used to test the sensitivity of the parameters to the number of sensors and their geometry, and to test if merging of the data could be of practical interest for the routine seismicity studies. The moment magnitude of the events varied between -1.4 and 0.9. In total 962 waveforms were analyzed. The results of the tests for the 30 studied events showed that the source locations of seismic event in the individual datasets (local and permanent) were not exactly the same and the differences between the locations could be up to 300 m depending on the azimuthal sensor coverage. In case of good sensor coverage, the hypocentral locations were more accurate and the difference between the locations obtained by each seismic system dataset was smaller than in case of bad sensor coverage. Some dynamic source parameters, on the other hand, were less affected by sensor coverage. However, it was found that a few dynamic parameters as energy, energy ratio (Es/Ep), apparent stress and stress drop were very sensitive to the geometry of the seismic system and change of up to 300% were observed in case of using different sensors datasets from sensors with bad coverage. The local network seismic energy was found to be always higher than the seismic energy obtained by the data from the permanent system independently on the sensor coverage. The consistent difference was attributed to underestimated correction for the attenuation, which affects more the results obtained from the permanent system with sensors generally at larger distances. A recommendation was made that the merged database should be used to obtain more accurate results, especially for hypocenter location, as the sensors of both systems together provide the best possible coverage but merging the data for calculation of dynamic parameters should be used with caution. In conclusion, the results in this thesis can be divided into two parts: first, results related to the reaction of the rock mass to blasting for Zinkgruvan and Garpenberg Mine, and second, methodological results related to the sensitivity of the kinematic and dynamic parameters to the number and geometry of the sensor locations, and practical application for Garpenberg Mine. Valuable information was obtained also about the orientation of the cluster of seismicity after the damaging event on July 2, 2015 at Zinkgruvan mine. This information can be useful to study the relationship between the damage and the source parameters and ultimately to improve the understanding the damage potential from seismic events in this the mines. All results could be used for further investigation on re-entry protocols for these two mines and seismic hazard assessment.